[MLA] Spotify’s plans

[Grimm | Eelco]

Hi Ian,

> I heard from somewhere rumours that in future Spotify will be normalising before encoding. 
> 
> (I can’t remember where this info came from, or whether it was on the record, which is why I’m mentioning it here first rather than in one of the AES groups)

Yes, we should keep this ‘under the hat’ for a while. I can confirm it is an idea that Spotify is considering. I have shared the idea with people from the target level group since I believe it should be part of our recommendation. I also mentioned that Spotify is considering it but that it was not final and should be treated as confidential info.

> Ultimately I think this is a good idea, since it will reduce the chances of decoder clipping, and also give better encoding results in general.

That is exactly why I told Spotify it is a good idea.

> However it’s also concerning if they choose (for example) -14 LUFS as a reference level. There will be a huge investment of time and processing in re-encoding their entire library, and I  imagine this would make them highly resistant to future changes - for example, to a reference level of -18 LUFS, as the AES are discussing.

No, it will always be a hybrid solution, which means: 1. attenuate PCM audio to -11 LUFS before encoding (the loudest target). 2. add metadata.

> Did this info come from anyone in this group ? If so, do we need to start a conversation about it, or are Spotify already waiting to see what the AES recommend ?

They absolutely do. One more reason to keep moving forward.

> If we don’t know, do you think it’s OK for me to post the same question in the AES groups ?!

No, we should not post our knowledge about Spotify’s plans to the larger AES group. We can post the idea of attenuating to -11 LUFS before encoding, but I rather first discuss it in the target group.

Cheers,
Eelco